Brain, Behavior, and Immunity 19 (2005) 275–280

www.elsevier.com/locate/ybrbi

Invited minireview

Chronic stress and comfort foods: Self-medication

and abdominal obesity夽
Mary F. Dallman¤, Norman C. Pecoraro, Susanne E. la Fleur
Department of Physiology and Program in Neurosciences, University of California San Francisco, San Francisco, CA 94143-0444, USA

Received 28 October 2004; received in revised form 11 November 2004; accepted 11 November 2004
Available online 29 January 2005

Abstract

Central corticotropin-releasing factor (CRF) networks are recruited by chronic stressors and elevated glucocorticoids (GCs) that
initiate recruitment of central CRF activity in the amygdala. Increased central activity of the CRF network stimulates all monoamin-
ergic cell groups, as well as premotor autonomic and other limbic structures resulting in the typical arousal, behavioral changes,
autonomic, and neuroendocrine changes that accompany the chronic imposition of a stressor. By contrast, elevated GCs appear,
through a variety of means to counteract the eVects of central CRF, which they have initiated. Together with insulin, the GCs stimu-
late drive for and ingestion of “comfort foods” that may directly result in reduction of the negative eVects of the chronic stressor in
the nucleus Accumbens, through stimulation of the anterior, more pleasure-associated part of this cell group, thus reducing the
weight of the stress-stimulated posterior, more defensive part. Furthermore, the shift in caloric intake from chow to preference for
“comfort foods,” together with elevated GCs and insulin, reorganize energy stores from a peripheral to a central distribution, pri-
marily as abdominal fat. A signal associated with this fat depot appears, as with eating “comfort foods,” to reduce the inXuence of
the chronic stress network on behaviors, autonomic, and neuroendocrine outXow.
 2004 Elsevier Inc. All rights reserved.

1. Introduction components of the system from the hypothalamus (cor-

The hypothalamo–pituitary–adrenal (HPA) axis is a
prototype of neuroendocrine systems with inhibitory
feedback loops mediated by the hormone secreted from
a remote target gland. The HPA axis is stimulated by
stressors (both physical and psychological insults), and
exhibits a strong circadian rhythm that peaks just prior
to the onset of the daily activity cycle. Its activity is
inhibited by adrenal glucocorticoid secretion, in a “long-
loop feedback.” In the 1950s, with new capability to
measure circulating corticosteroid concentrations in
humans and animals, it became abundantly clear that
treatment with glucocorticoids (GC) dramatically
reduced the secretory capacity of the forward, motor
夽
Supported, in part, by NIH Grants DK28172 and DA16944.
*
Corresponding author. Fax: +1 415 476 4929.
E-mail address: dallman@itsa.ucsf.edu (M.F. Dallman).
ticotropin-releasing factor; CRF) and from the anterior
pituitary (adrenocorticotropin; ACTH). This recogni-
tion was in large part responsible for the routine extra
‘surgical coverage’ for individuals treated with life-sav-
ing GCs for other disease processes (Munck et al., 1984).
However, most of the early experimental studies exam-
ined the eVect of prior treatment with exogenous GCs on
subsequent basal and stress-induced activity in the HPA
axis. Under basal conditions, prolonged treatment with
GCs does, indeed, inhibit most components of the HPA
axis (Dallman et al., 1987).
However, when tests were made of the eVects of
stressor-induced, endogenous GC secretion, subsequent
CRF and ACTH responses to novel stressors were not
inhibited, although stress-induced GC secretion was
suYcient to inhibit ACTH secretion to acute stress in the
absence of concurrent stress (Dallman et al., 1987). GC
feedback was ineVective in the presence of a stressor, and

0889-1591/$ - see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bbi.2004.11.004
276 M.F. Dallman et al. / Brain, Behavior, and Immunity 19 (2005) 275–280

high concentrations were required in chronically stressed
animals for a normal HPA response to a novel stressor.
Exploration of the relationships among acute and
chronic stressors, the brain and the HPA axis shows the
following: (1) GCs act in a direct, feedback inhibitory
mode to reduce the duration of any single episode of
CRF and ACTH secretion; (2) GCs act directly in a
feed-forward, excitatory or positive, mode on brain to
increase drive, and CRF and ACTH secretion; (3) in the
presence of insulin, stress, and GCs increase the relative
intake of “comfort foods” and reduce HPA activity;
and, (4) GCs act on peripheral energy storage organs to
promote central obesity, providing an indirect, second-
ary inhibitory feedback signal that limits the degree of
chronic stress perceived by the organism. Here we brieXy
review these eVects and interactions of stress, GCs, drive,
and metabolism.
1.1. GCs act in a direct, feedback inhibitory mode to
reduce the duration of any single episode of CRF and
ACTH secretion (Fig. 1, bottom left)
The ACTH response to an acute stressor is tempo-
rally limited is greatest shortly after the onset of the
stressor and may return to basal concentrations during
the sustained application of a mild stressor, such as tube-
restraint, provided that GCs can be secreted endoge-
nously or are infused. However, if the adrenals cannot
secrete GCs and the steroids are not infused, ACTH
responses to a stimulus are prolonged, and may persist
for hours. The normal, acute rise in GCs that follow an
ACTH response to a stressor acts on CRF neurons to
inhibit aVerent inputs in both rats and rhesus monkeys
(Keller-Wood and Dallman, 1984). The eVect is both at
the hypothalamus and pituitary corticotropes and
appears to be a rapid, non-genomically mediated action
of the GCs. A recent study from Tasker’s lab showed in
hypothalamic slices that dexamethasone infusion acti-
vated cannabinoid secretion from CRF neurons; the
cannabinoids acted locally on CB-1 cannabinoid recep-
tors on aVerent glutamatergic axon terminals to block
input to the CRF cells (Di et al., 2003). This provides an
excellent mechanism for the rapid eVects of GCs on
activity of CRF neurons, and similar mechanisms may
occur elsewhere in brain and at the corticotropes. This
fast negative feedback reins in secretion of GCs to a level
of suYciency, restraining them from a potentially dam-
aging excess.
1.2. GCs act directly in a feed-forward, positive mode on
brain to increase food-associated drives, CRF and ACTH
secretion (Fig. 1, bottom left middle)
These unexpected Wndings came from experiments
that were designed to test other propositions. It is widely
agreed that both stress and glucocorticoid treatment
aVect responses to drugs of abuse (Goeders, 2002; Koob
and Le Moal, 2001). We used pleasurable, high-density
sucrose (30%) as a drug-surrogate to test the eVect of
varied concentrations of corticosterone in adrenalecto-
mized rats on the amount of sucrose drunk (Bell et al.,
2000); we also tested the eVect of corticosterone on sac-
charin intake (Bhatnagar et al., 2000). The amounts of
both sweet drinks ingested by adrenalectomized rats
depended on circulating corticosterone. Furthermore,
adrenalectomized rats respond to corticosterone in a
dose-related manner to engage in running-wheel behav-
ior (Leshner, 1971) and in schedule-induced polydipsia
(Cirulli et al., 1994), both of which are motivated behav-
iors surrounding food intake.
The corticosterone dose-related responses above are
probably associated with the eVects of corticosterone on

Fig. 1. Stress and glucocorticoids (GCs) act on brain to alter activity in various sites with varying eVects. Bottom left to right: GC fast-feedback acts
at the hypothalamus (PVN) and corticotrope cells of the anterior pituitary (a. Pit) to inhibit the duration of ACTH secretion in response to acute
stress. GCs, in the presence of insulin, stimulate motivation for “comfort foods,” an action probably exerted in the meso-limbic dopaminergic system
(VTA, ACC shell). Stress and elevated GCs activate the central CRF chronic stress network; high glucocorticoids stimulate CRF expression in the
limbic central nucleus of the amygdala (CeA). Peripheral actions of GCs and insulin redistribute energy stores to central fat; a signal from abdominal
fat stores serves to inhibit CRF activity in the paraventricular hypothalamus (PVN).
 M.F. Dallman et al. / Brain, Behavior, and Immunity 19 (2005) 275–280
the meso-limbic dopaminergic brain “reward” system.
Adrenalectomy decreases dopamine release speciWcally
in the shell of the nucleus Accumbens (n. Accsh) in
response to both drug injections and hypothalamic self-
stimulation, and treatment with corticosterone restores
both to normal (Barr et al., 2000; Goeders, 2002). More-
over dopamine transporters in the n. Accsh are reduced
by adrenalectomy and restored in a dose-related fashion
by corticosterone treatment (Sarnyai et al., 1998). Thus,
responsivity to many stimuli is altered by the GC milieu
apparently acting on the dopaminergic system in n.
Accsh. Taken together, these results suggest strongly that
corticosterone may generally increase food-associated
‘drives,’ stimulus salience or motivation, possibly
through augmented dopamine secretion in the n. Accsh.
But, corticosterone does not appear to increase feed-
ing-motivated behaviors under all conditions. Although
adrenalectomized rats eat slightly less chow than con-
trols, and corticosterone restores chow intake to normal,
there is not a corticosterone dose-related increase of
chow intake in adrenalectomized rats. The drive to
obtain food is one of the most potent that we have, and
this observation appeared to negate the drive-promoting
eVect that we attempted to ascribe to corticosterone.
However, when rats are made diabetic with streptozoto-
cin (that kills pancreatic B-cells, and thus insulin secre-
tion), a marked, dose-dependent eVect of corticosterone
on intake of rat chow is revealed. A major diVerence
between adrenalectomized and adrenalectomized-dia-
betic rats is the lack of insulin; moreover, circulating
insulin concentrations increase directly as a function of
circulating corticosterone concentrations in adrenalecto-
mized rats (Dallman et al., 2002).
In a recent series of experiments, we have retested
the eVects of increasing corticosterone on chow intake
and tested in similar experiments the eVects of
increasing steroid on lard ingestion in adrenalecto-
mized rats (la Fleur et al., 2004). The results of these
experiments provide a possible explanation for the
nonlinear eVects of corticosterone on chow intake. As
previously, increasing corticosterone concentrations
increased insulin in adrenalectomized rats, and did not
increase chow intake; however, chow intake was
increased in a corticosterone dose-dependent fashion in
adrenalectomized-diabetic rats. By contrast, increasing
corticosterone concentrations did increase lard intake
in a dose-related fashion in adrenalectomized rats but
did not aVect lard intake in adrenalectomized-diabetic
rats. Finally, when adrenalectomized-diabetic rats
treated with high corticosterone were also infused sub-
cutaneously with low doses of insulin, lard intake was
directly proportional to circulating insulin concentra-
tions in the diabetic rats (la Fleur et al., 2004). The
results of these experiments suggest strongly that corti-
costerone is, indeed, a hormone that generally increases
food-associated drives or motivation; it may be that
277
insulin determines the preference for which foods are
ingested.
1.3. Chronic stress recruits CRF systems in brain (Fig. 1,
bottom right middle), but stress and GCs also increase the
relative intake of “comfort foods” that damp stress
responses
Chronically stressed rats persistently hypersecrete
glucocorticoids during application of the stressor, com-
pared to unstressed controls. Characteristically, if sub-
jected to a new stressor during a period of chronic stress,
hypothalamic CRF, ACTH, and corticosterone
responses are increased, compared to control, as are
autonomic and behavioral responses (Dallman et al.,
2002). These responses are probably a consequence of
recruitment of a central stress response network (Dall-
man et al., 2002) regulated by increased activity in the
central, non-hypothalamic CRF system (Schulkin et al.,
1994), and induced by an action of corticosterone at the
amygdala (Shepard et al., 2000). Thus, during chronic
stress, corticosterone acts in a feed-forward, positive
fashion to augment acute ACTH secretion while still
limiting its duration through the rapid negative feedback
eVect. When the actual stress episode is over, systemic
activity in the HPA axis (ACTH and GCs) may be
remarkably subnormal, as is seen in patients with post-
traumatic stress disorders (Aardal-Eriksson et al., 2001)
and in rats after withdrawal from morphine (Houshyar
et al., 2004).
Because it seems clear that in the presence of insulin
elevated glucocorticoids increase drive for pleasurable
foods, we tested whether exposing intact rats to repeated
restraint stress altered the amount of “comfort foods”
they ingested (Pecoraro et al., 2004). Rats were provided
with dishes of lard and bottles of 30% sucrose as well as
chow for three day to remove any neophobia for these
foods before they were exposed to restraint. After a day
without these “comfort foods,” in a 2 £ 2 design, half of
the rats were resupplied with ad lib access to “comfort
foods” and half were exposed to 3 h/day tube restraint
stress for Wve consecutive days.
Because of the prior ingestion of “comfort foods” the
two groups eating these were fatter. Generally, adult
male rats reduce chow intake and stop growing during
periods of sustained or repeated stress (Dallman and
Bhatnagar, 2001); eating “comfort foods” during
repeated restraint allowed this group both ponderal
growth and increased caloric eYciency during chronic
stress. The increased caloric eYciency (g gained/cal
ingested) suggests that the normal degree of stress-
induced sympathetic neural outXow was reduced in the
stressed group eating “comfort foods.” Rats that ate
“comfort foods” had reduced HPA responses to
restraint and lower basal CRF in the hypothalamus
(Pecoraro et al., 2004).
278 M.F. Dallman et al. / Brain, Behavior, and Immunity 19 (2005) 275–280
In another set of experiments, we held corticosterone
constant at normal and stress concentrations after adre-
nalectomy, and exposed half of the rats to cold stress
§30% sucrose to drink (Bell et al., 2002). Again, “com-
fort food” reduced the degree of metabolic stress
induced by cold, suggesting strongly that “comfort
food” reduces the degree of stressor-induced sympa-
thetic responses.
In both of the above experiments, provision of “com-
fort foods” reduced the magnitude of HPA responses to
the stressors (Bell et al., 2002; Pecoraro et al., 2004). Sim-
ilar inhibition of the HPA axis occurs when a variable
stress paradigm is applied to obesity-sensitive, but not
obesity-resistant rats eating a high-fat diet (Levin et al.,
2000). Thus, it appears that provision of “comfort
foods” either as fat, high sucrose concentrations, or the
combination reduces both autonomic and HPA
responses to repeated stressors in rats.
There is some evidence that “comfort foods” given to
chronically stressed rats may negate chronic stress-
induced inhibition of dopamine release that occurs in the
n. Accsh. Although acute stress stimulates dopamine
secretion in both n. Accsh and prefrontal cortex (Ber-
ridge, 2004), chronic stress inhibits dopamine secretion
in these sites [see (Nanni et al., 2003)]. A learned appeti-
tive behavior in combination with chronic stress pre-
vented the inhibitory eVects of a chronic stress procedure
on dopamine output and dopamine transporter density
in n. Accsh (Nanni et al., 2003). Moreover, brief sched-
uled presentation of palatable sucrose increases chow
intake, and dopamine transporters in the ventral teg-
mental area and n. Accsh of rats on a restricted, but not
ad lib chow diet (Bello et al., 2003). Restricted feeding
alters HPA activity so that it is maximal just prior to
expected caloric intake (Pecoraro et al., 2002). Thus, it
may be that the inhibition of HPA responses to stressors
in rats eating “comfort foods” is explained by the inter-
play between the negative eVects of chronic stressors,
and the positive eVects of “comfort foods” on inputs to
the VTA-n. Accsh reward network.
1.4. But GCs also act on peripheral energy storage organs
to promote central obesity (Fig. 1, bottom right)
In the presence of insulin, passive treatment of rats
with high GCs reduces chow intake, body weight, sym-
pathetic activity but increases fat stores (Dallman et al.,
2002; la Fleur et al., 2004). With chronic stressors and
endogenous corticosterone, there is also decreased chow
intake, body weight and a relative increase in central fat
(Houshyar et al., 2004). This is a result of the action of
GCs in mobilizing small molecules (amino acids and
fatty acids) from peripheral protein (muscle, skin) and
fat stores for use in hepatic gluconeogenesis and keto-
genesis. In the presence of insulin and high GCs, fat
accrues in intra-abdominal stores, where lipids serve,
upon mobilization, as immediate further substrate for
hepatic gluconeogenesis as well as ketones for direct
energy use by brain (Peters et al., 2004).
Intra-abdominal fat stores appear to serve as an
excellent surrogate for an indirect metabolic negative
feedback signal of glucocorticoids since they are highly
negatively correlated with hypothalamic CRF expres-
sion and HPA responsivity in both adrenalectomized-
corticosterone replaced, and in intact, chronically
stressed rats (Dallman et al., 2003; Houshyar et al.,
2004). The speciWc signal to brain that represents
increased abdominal fat stores is still unidentiWed. How-
ever, it seems clear that this signal does act at brain to
reduce the adverse eVects of a recruited chronic stress
response network, and probably makes animals (and
people) feel better under conditions of chronic stress.
1.5. Summary (Fig. 1) and perspective
Central CRF networks are recruited by chronic
stressors and elevated GCs initiating recruitment of cen-
tral CRF activity in the amygdala. Increased central
activity of the CRF network stimulates all monoaminer-
gic cell groups, as well as premotor autonomic and other
limbic structures resulting in the typical arousal, behav-
ioral changes, autonomic, and neuroendocrine changes
that accompany the chronic imposition of a stressor. By
contrast, elevated GCs appear, through a variety of
means, to counteract the eVects of central CRF, which
they have initiated. Together with insulin, the GCs stim-
ulate drive for and ingestion of “comfort foods” that
may directly result in reduction of the negative eVects of
the chronic stressor in the n. Accsh, through stimulation
of the anterior, more pleasure-associated part of this cell
group, thus reducing the weight of the stress-stimulated
posterior, more defensive part (Berridge, 2004). Further-
more, the shift in caloric intake from chow to preference
for “comfort foods,” together with elevated GCs and
insulin, reorganize energy stores from a peripheral to a
central distribution, primarily as abdominal fat. A signal
associated with this fat depot appears, as with eating
“comfort foods” to reduce the inXuence of the chronic
stress network on behaviors, autonomic, and neuroen-
docrine outXow.
There is evidence that this sequence of events of
chronic stressors on rats also may apply to humans. Stu-
dents under stress report shifting ingestion from normal
(fruits, vegetables, Wsh, and meat) to sweet and savory
foods (Oliver and Wardle, 1999). Moreover, stress pre-
cipitates binge eating (Freeman and Gil, 2004; Schoe-
maker et al., 2002), and women under acute lab stress
increase “comfort food” intake (Epel et al., 2001).
Unfortunately, persisting in this behavior either as habit
or for self-medication yields abdominal obesity.
Increased abdominal fat stores are strongly associated
with the metabolic syndrome, hypertension, type 2
 M.F. Dallman et al. / Brain, Behavior, and Immunity 19 (2005) 275–280
diabetes, and cardiovascular disease, morbidity and
mortality (Friedman, 2003; Stunkard et al., 2003).
Rather than indulging in “comfort foods” to feel bet-
ter during periods of chronic stress, it behooves us to
Wnd a means to remove or reduce the chronic stressor. A
recent evaluation of the eVects of a program designed to
remove impoverished people from inner city public
housing where the poverty level was initially >40% to
sites with far less poverty showed signiWcant self-
reported eVects that included both reduced sensations of
stress, increased peacefulness, and weight loss, as com-
pared with a randomly assigned control group who
remained in high-poverty areas (Orr et al., 2003). Per-
haps it would help to relieve some of the current epi-
demic of obesity were policy-makers aware of the
insidious eVects of not actively seeking to relieve sources
of uncontrollable chronic stressors in our current lives.
References
Aardal-Eriksson, E., Eriksson, T., Thorell, L.-H., 2001. Salivary corti-
sol, posttraumatic stress symptoms, and general health in the acute
phase and during a 9-month follow-up. Biol. Psychiatry 50, 986–
993.
Barr, A.M., Brotto, L.A., Phillips, A.G., 2000. Chronic corticosterone
enhances the rewarding eVect of hypothalamic self-stimulation in
rats. Brain Res. 875, 196–201.
Bell, M.E., Bhatnagar, S., Liang, J., Soriano, L., Nagy, T.R., Dallman,
M.F., 2000. Voluntary sucrose ingestion, like corticosterone
replacement, prevents the metabolic deWcits of adrenalectomy. J.
Neuroendocrinol. 12, 461–470.
Bell, M.E., Bhargava, A., Soriano, L., Laugero, K., Akana, S.F.,
Dallman, M.F., 2002. Sucrose and corticosterone interact to mod-
ulate behavior, energy balance, autonomic outXow and neuroen-
docrine responses during chronic cold. J. Neuroendocrinol. 14,
330–342.
Bello, N.T., Sweigart, K.L., Lakoski, J.M., Norgren, R., Hajnal, A.,
2003. Restricted feeding with scheduled sucrose access results in an
upregulation of the rat dopamine transporter. Am. J. Physiol. 284,
R1260–R1268.
Berridge, K.C., 2004. Motivation concepts in behavioral neuroscience.
Physiol. Behav. 81, 179–209.
Bhatnagar, S., Bell, M.E., Liang, J., Soriano, L., Nagy, T.R., Dallman,
M.F., 2000. Corticosterone facilitates saccharin intake in adrenalec-
tomized rats. Does corticosterone increase stimulus salience. J.
Neuroendocrinol. 12, 453–460.
Cirulli, F., van Oers, H., de Kloet, E.R., Levine, S., 1994. DiVerential
inXuence of corticosterone and dexamethasone on schedule-
induced polydipsia in adrenalectomized rats. Behav. Brain Res. 65,
33–39.
Dallman, M.F., Akana, S.F., Cascio, C.S., Darlington, D.N., Jacobson,
L., Levin, N., 1987. Regulation of ACTH secretion: variations on a
theme of B. In: Clark, J.H. (Ed.), Recent Progress in Hormone
Research. Academic Press, Orlando, pp. 113–173.
Dallman, M.F., Bhatnagar, S., 2001. Chronic Stress and Energy Bal-
ance: Role of the Hypothalamo–Pituitary–Adrenal Axis. Hand-
book of Physiology, Section 7 The Endocrine System. Oxford
University Press, New York.
Dallman, M.F., Pecoraro, N., Akana, S.F., la Fleur, S.E., Gomez, F.,
Houshyar, H., Bell, M.E., Bhatnagar, S., Laugero, K.D., Manalo, S.,
2003. Chronic stress & obesity: a new view of comfort food. Proc.
Soc. Natl. Acad. Sci. USA 100, 11696–11701.
279
Dallman, M.F., Viau, V., Bhatnagar, S., Gomez, F., Laugero, K., Bell,
M.E., 2002. Corticotropin-releasing factor (CRF), corticosteroids,
stress and sugar: energy balance the brain and behavior. In: PfaV,
D.W. (Ed.), Hormones, Brain and Behavior. Academic Press, San
Diego, pp. 571–631.
Di, S., Malcher-Lopes, R., Halmos, K.C., Tasker, J.G., 2003. Nonge-
nomic glucocorticoid inhibition via endocannabinoid release in the
hypothalamus: a fast feedback mechanism. J. Neurosci. 23, 4850–
4857.
Epel, E., Lapidus, R., McEwen, B., Brownell, K., 2001. Stress may add
bite to appetite in women: a laboratory study of stress-induced cor-
tisol and eating behavior. Psychoneuroendocrinology 26, 37–49.
Freeman, L.M., Gil, K.M., 2004. Daily stress, coping and dietary
restraint in binge eating. Int. J. Eat. Disord. 36, 204–212.
Friedman, J.M., 2003. A war on obesity, not the obese. Science 299,
856–858.
Goeders, N.E., 2002. Stress and cocaine addiction. JPET 301, 785–789.
Houshyar, H., Manalo, S., Dallman, M.F., 2004. Time-dependent alter-
ations in mRNA expression of brain neuropeptides regulating
energy balance and hypothalamo–pituitary–adrenal activity after
withdrawal from intermittent morphine treatment. J. Neurosci. 24,
9414–9424.
Keller-Wood, M.E., Dallman, M.F., 1984. Corticosteroid inhibition of
ACTH secretion. Endocr. Rev. 5, 1–24.
Koob, G.F., Le Moal, M., 2001. Drug addiction, dysregulation of
reward, and allostasis. Neuropsychopharmacology 24, 97–129.
la Fleur, S.E., Akana, S.F., Manalo, S., Dallman, M.F., 2004. Interac-
tion between corticosterone and insulin in obesity: regulation of
lard intake and fat stores. Endocrinology 145, 2174–2185.
Leshner, A.I., 1971. The adrenals and the regulatory nature of running
wheel activity. Physiol. Behav. 6, 551–558.
Levin, B.E., Richard, D., Michel, C., Servatius, R., 2000. DiVerential
stress responsivity in diet-induced obese and resistant rats. Am. J.
Physiol. 279, R1357–R1364.
Munck, A., Guyre, P.M., Holbrook, N.J., 1984. Physiological functions
of glucocorticoids in stress and their relation to pharmacological
actions. Endocr. Rev. 5, 25–44.
Nanni, G., Scheggi, S., Leggio, B., Grappi, S., Masi, F., Rauggi, R., De
Montis, M.G., 2003. Acquisition of an appetitive behavior prevents
development of stress-induced neurochemical modiWcations in rat
nucleus accumbens. J. Neurosci. Res. 73, 573–580.
Oliver, G., Wardle, J., 1999. Perceived eVects of stress on food choice.
Physiol. Behav. 66, 511–515.
Orr, L., Feins, J.D., Jacob, R., Beecroft, E., Sanbonmatsu, L., Katz,
L.F., Liebman, J.B., Kling, J.R., 2003. Moving to opportunity
interim impacts evaluation. Final Report. US Department of Hous-
ing and Urban Development; OYce of Policy Development and
Research. pp. 1–165.
Pecoraro, N., Gomez, F., Laugero, K., Dallman, M.F., 2002. Brief
access to sucrose engages food-entrainable rhythms in food-
deprived rats. Behav. Neurosci. 116, 757–776.
Pecoraro, N., Reyes, F., Gomez, F., Bhargava, A., Dallman, M.F., 2004.
Chronic stress promotes palatable feeding, which reduces signs of
stress: feedforward and feedback eVects of chronic stress. Endocri-
nology 145, 3754–3762.
Peters, A., Schweiger, U., Pellerin, L., Hubold, C., Oltmanns, K.M.,
Conrad, M., Schultes, B., Born, J., Fehm, H.L., 2004. The selWsh
brain: competition for energy resources. Neurosci. Biobehav. Rev.
28, 143–180.
Sarnyai, Z., McKitterick, C.R., McEwen, B.S., Kreek, M.J., 1998. Selec-
tive regulation of dopamine transporter binding in the shell of the
nucleus accumbens by adrenalectomy and corticosterone-replace-
ment. Synapse 30, 334–337.
Schoemaker, C., Smit, F., Bijl, R.V., Vollebergh, W.A.M., 2002. Bulimia
nervosa following psychological and multiple child abuse: support
for the self-medication hypothesis in a population-based cohort
study. Int. J. Eat. Disord. 32, 381–388.
280 M.F. Dallman et al. / Brain, Behavior, and Immunity 19 (2005) 275–280

Schulkin, J., McEwen, B.S., Gold, P.W., 1994. Allostasis, amygdala, and in the central amygdaloid nucleus and anxiety-like behavior. Brain
anticipatory angst. Neurosci. Biobehav. Rev. 18, 385–396. Res. 861, 288–295.
Shepard, J.D., Barron, K.W., Myers, D.A., 2000. Corticosterone deliv- Stunkard, A.J., Faith, M.S., Allison, K.C., 2003. Depression and obes-
ery to the amygdala increases corticotropin-releasing factor mRNA ity. Biol. Psychiatry 54, 330–337.